Z.ai: GLM 4.5V

Text input Image input Text output
Author's Description

GLM-4.5V is a vision-language foundation model for multimodal agent applications. Built on a Mixture-of-Experts (MoE) architecture with 106B parameters and 12B activated parameters, it achieves state-of-the-art results in video understanding, image Q&A, OCR, and document parsing, with strong gains in front-end web coding, grounding, and spatial reasoning. It offers a hybrid inference mode: a "thinking mode" for deep reasoning and a "non-thinking mode" for fast responses. Reasoning behavior can be toggled via the `reasoning` `enabled` boolean. [Learn more in our docs](https://openrouter.ai/docs/use-cases/reasoning-tokens#enable-reasoning-with-default-config)

Key Specifications
Cost
$$$$
Context
65K
Parameters
106B (Rumoured)
Released
Aug 11, 2025
Speed
Ability
Reliability
Supported Parameters

This model supports the following parameters:

Tools Temperature Max Tokens Top P Tool Choice Reasoning Include Reasoning
Features

This model supports the following features:

Tools Reasoning
Performance Summary

Z.AI's GLM-4.5V demonstrates moderate speed performance, ranking in the 30th percentile across benchmarks, and offers moderate pricing, placing it in the 25th percentile. A standout feature is its exceptional reliability, boasting a 98% success rate, indicating minimal technical failures and consistent response delivery. In terms of accuracy, GLM-4.5V excels in several areas. It achieves strong results in Coding (92.0% accuracy, 79th percentile), Email Classification (99.0% accuracy, 80th percentile), and Reasoning (92.0% accuracy, 82nd percentile), aligning with its description as a vision-language model for multimodal agent applications with strong gains in grounding and spatial reasoning. General Knowledge (99.0% accuracy, 67th percentile) and Ethics (99.0% accuracy, 54th percentile) also show solid performance. Its hallucination rate is relatively low at 92.0% accuracy, suggesting a good ability to acknowledge uncertainty. However, a notable weakness is its Instruction Following capability, with a significantly lower accuracy of 5.1% (23rd percentile), indicating challenges with complex multi-step directives. Mathematics performance is average at 85.0% accuracy (52nd percentile).

Model Pricing

Current Pricing

Feature Price (per 1M tokens)
Prompt $0.6
Completion $1.8
Input Cache Read $0.11

Price History

Available Endpoints
Provider Endpoint Name Context Length Pricing (Input) Pricing (Output)
Z.AI
Z.AI | z-ai/glm-4.5v 65K $0.6 / 1M tokens $1.8 / 1M tokens
Novita
Novita | z-ai/glm-4.5v 65K $0.6 / 1M tokens $1.8 / 1M tokens
Parasail
Parasail | z-ai/glm-4.5v 65K $0.6 / 1M tokens $1.8 / 1M tokens
DeepInfra
DeepInfra | z-ai/glm-4.5v 65K $0.6 / 1M tokens $1.8 / 1M tokens
Benchmark Results
Benchmark Category Reasoning Strategy Free Executions Accuracy Cost Duration
Other Models by z-ai